Wednesday, 2017-03-22

*** ganso has quit IRC00:25
*** kaisers_ has joined #openstack-manila00:26
*** kaisers_ has quit IRC00:31
*** crushil has joined #openstack-manila00:41
*** harlowja has joined #openstack-manila00:42
tommylikehuping tbarron00:59
*** ganso has joined #openstack-manila00:59
tommylikehuevening!,  anyone have any idea what is Jan Provaznik's irc name?01:02
bswartztommylikehu: jprovazn01:05
tommylikehubswartz: thanks01:05
tommylikehubswartz: do you know when is he ususally around?01:06
bswartzhe's based in Europe01:06
bswartzso I imagine it's 2AM or 3AM where he is01:06
bswartzhe'll probably wake up in 4-5 hours01:06
tommylikehuwow, I will try to contact him after lunch, thanks.01:08
*** markstur has quit IRC01:16
*** kaisers_ has joined #openstack-manila01:27
*** kaisers_ has quit IRC01:31
*** gouthamr has joined #openstack-manila01:32
*** tongli has joined #openstack-manila01:33
*** kaisers1 has joined #openstack-manila01:34
*** kaisers has quit IRC01:35
*** markstur has joined #openstack-manila01:39
*** cknight has joined #openstack-manila02:03
*** tongli has quit IRC02:26
*** kaisers has joined #openstack-manila02:28
*** kaisers has quit IRC02:33
*** gouthamr has quit IRC02:45
*** cknight has quit IRC02:54
*** faizy has joined #openstack-manila03:34
*** harlowja has quit IRC03:35
*** faizy has quit IRC03:39
*** kaisers has joined #openstack-manila04:00
*** kaisers has quit IRC04:05
*** swapnil-nilangek has joined #openstack-manila04:08
*** gouthamr has joined #openstack-manila04:25
*** ganso has quit IRC04:35
*** Administrator_ has joined #openstack-manila04:38
*** crushil has quit IRC04:38
*** zhugaoxiao has quit IRC04:41
*** kaisers has joined #openstack-manila04:44
*** gcb has quit IRC04:53
*** gouthamr has quit IRC05:02
*** arnewiebalck_ has joined #openstack-manila05:05
*** guchihiro has joined #openstack-manila05:11
*** gcb has joined #openstack-manila05:11
*** guchihiro has quit IRC05:12
*** arnewiebalck_ has quit IRC05:14
*** faizy has joined #openstack-manila05:19
*** faizy has quit IRC05:22
*** faizy has joined #openstack-manila05:22
*** faizy_ has joined #openstack-manila05:23
*** faizy_ has quit IRC05:27
*** faizy_ has joined #openstack-manila05:27
*** faizy has quit IRC05:27
*** faizy_ has quit IRC05:30
*** faizy_ has joined #openstack-manila05:31
*** lpetrut has joined #openstack-manila05:38
*** kaisers has quit IRC05:49
*** faizy_ has quit IRC06:07
*** faizy has joined #openstack-manila06:08
*** faizy has quit IRC06:11
*** faizy has joined #openstack-manila06:11
*** arnewiebalck_ has joined #openstack-manila06:12
*** arnewiebalck_ has quit IRC06:13
*** faizy has quit IRC06:15
*** faizy has joined #openstack-manila06:15
*** faizy has quit IRC06:18
*** faizy has joined #openstack-manila06:19
*** lpetrut has quit IRC06:19
*** faizy has quit IRC06:21
*** faizy has joined #openstack-manila06:22
*** markstur has quit IRC06:25
*** nkrinner_afk is now known as nkrinner06:43
*** kaisers has joined #openstack-manila06:50
*** kaisers has quit IRC06:54
*** dsariel_ has joined #openstack-manila07:02
*** jprovazn has joined #openstack-manila07:28
*** markstur has joined #openstack-manila07:39
*** pcaruana has joined #openstack-manila07:39
*** markstur has quit IRC07:43
*** guchihiro has joined #openstack-manila07:50
*** chenying has quit IRC07:51
*** chenying has joined #openstack-manila07:52
openstackgerritTina Tang proposed openstack/manila master: [Dell EMC Unity] Create with user capacity  https://review.openstack.org/44842907:53
*** faizy has quit IRC07:54
*** markstur has joined #openstack-manila08:04
*** swapnil-nilangek has quit IRC08:06
*** markstur has quit IRC08:08
*** kaisers1 has quit IRC08:09
*** kaisers has joined #openstack-manila08:10
*** swapnil-nilangek has joined #openstack-manila08:25
*** guchihiro has quit IRC08:25
*** markstur has joined #openstack-manila08:28
*** rraja has joined #openstack-manila08:32
*** markstur has quit IRC08:33
openstackgerritzhongjun proposed openstack/manila master: change user access name limit from 32 to 255 characters  https://review.openstack.org/44844608:43
*** arnewiebalck_ has joined #openstack-manila08:46
tommylikehuhey jprovazn, are you around?08:49
jprovazntommylikehu, hi, yes08:49
*** kaisers_ has joined #openstack-manila08:51
tommylikehujprovazn: I wanna talk about this patch. https://review.openstack.org/#/c/443101/ I add some comments few days ago08:52
jprovazntommylikehu, thanks for looking at it. Let me check your comments08:54
tommylikehunp, I think we are trying to catch and classify the exceptions and what is exactly already did by exception, So I am wondering if we can add the exception into the event_ids08:54
tommylikehuwith that we don't need that many event_ids and no need to classify these exceptions either08:55
jprovazntommylikehu, providing more details about the error is tricky - user rmessages intentionally avoid providing any details because of potential exposing of security information08:57
*** kaisers_ has quit IRC08:57
openstackgerritzhongjun proposed openstack/python-manilaclient master: change user access name limit from 32 to 255 characters  https://review.openstack.org/44845008:57
jprovazntommylikehu, yes, more detailed exception classification would be better, though we don't do this in code so far08:58
jprovazntommylikehu, though this patch might be good start for better categorization if you mean this :)08:58
tommylikehujprovazn that could be a good start08:59
tommylikehuso the only reason that prevent us from doing that is the worries of exposing security information?08:59
tommylikehuI have another concern about this kind of implementation. I just uploaded another draft patch whichi is doing what I am thinking09:00
tommylikehuhttps://review.openstack.org/#/c/448384/09:00
tommylikehujprovazn: would you take a look and leave some comments?09:01
jprovazntommylikehu, yes, looking09:02
*** ociuhandu has quit IRC09:03
tommylikehujprovazn: thanks, but I have to say I didn't considered the security issues09:03
jprovazntommylikehu, https://specs.openstack.org/openstack/cinder-specs/specs/newton/summarymessage.html#security-impact - I can imagine that just exposing backend driver error to end user may be undesired by admins09:08
tommylikehujprovazn: if that is one of the most concerns I think it's easy to fix, just build the relationship from exception to visible user messages09:12
jprovazntommylikehu, I think there are 2 questions: 1) could we expose more info about error to end-user (your comment https://review.openstack.org/#/c/443101/5/manila/message/defined_messages.py) - we have to consider a security concern here. 2) could we classify errors based on exception type - yes, we could and it would be probably cleaner, but it requires more code refactoring I guess09:15
tommylikehujprovazn: yes, maybe you are right, but if we gonna add more user messages in this way: https://github.com/openstack/cinder/blob/master/cinder/volume/manager.py#L137109:17
tommylikehujprovazn: I think it's requiring more codes09:18
tommylikehujprovazn: and more complicated to maintaIN09:18
jprovazntommylikehu, regarding (2) - yes, nicer solution would be raising classified exception for each error and then just provide a fance user message based on this. But how share service code typically looks like now is:09:21
jprovaznhttps://github.com/openstack/manila/blob/master/manila/share/manager.py#L249709:21
jprovazntommylikehu, IOW if an excpetion during self.driver.revert_to_snapshot occurs, error state is set and another code is done after it09:22
tommylikehujprovazn: let me check the differencew09:23
tommylikehujprovazn: you mean you don't care what exactly happened in  the driver level?09:25
tommylikehus/you/manila09:26
*** arnewiebalck_ has quit IRC09:28
jprovazntommylikehu, my understanding is that your idea is to raise classified exceptions in share service and then in scheduler service create user message in one place based on the exception type - https://review.openstack.org/#/c/448384/3/cinder/scheduler/flows/create_volume.py09:35
jprovaznis my understanding correct?09:35
tommylikehujprovazn: yes09:35
tommylikehuuse exception to speak out the detail message09:36
jprovazntommylikehu, ok, I think it's good approach (on first I missed that we already have save_and_reraise_exception in all exception hadling so I thought that raising an exception would break code flow but I was wrong)09:38
jprovazntommylikehu, whether excpetion details should be exposed to end user is question for broad manila team I guess - maybe we could discuss this on tomorrow meeting09:40
openstackgerritTina Tang proposed openstack/manila master: [Dell EMC Unity] Create with user capacity  https://review.openstack.org/44842909:41
jprovazntommylikehu, I'll update my patch meantime09:41
tommylikehujprovazn: that's a nice advice. I already add the topic in cinder today, hope we could get more response from cinder guys09:44
tommylikehus/add/added09:45
jprovazntommylikehu, cool, let's see what they think09:46
*** gcb has quit IRC09:48
*** markstur has joined #openstack-manila09:55
*** dgonzalez_ has joined #openstack-manila09:55
*** databus23_ has joined #openstack-manila09:55
*** seife_ has joined #openstack-manila09:55
*** carthaca_ has joined #openstack-manila09:55
*** david_1 has joined #openstack-manila09:55
*** sapcc-bot1 has joined #openstack-manila09:55
*** seife_ has quit IRC09:57
*** databus23_ has quit IRC09:57
*** dgonzalez_ has quit IRC09:57
*** carthaca_ has quit IRC09:57
*** david_1 has quit IRC09:57
*** markstur has quit IRC09:59
*** kaisers_ has joined #openstack-manila10:53
*** kaisers_ has quit IRC10:58
*** markstur has joined #openstack-manila11:20
*** markstur has quit IRC11:25
*** gcb has joined #openstack-manila11:38
*** markstur has joined #openstack-manila11:45
*** markstur has quit IRC11:50
*** tommylikehu_ has joined #openstack-manila11:52
*** swapnil-nilangek has quit IRC11:58
*** sapcc-bot has quit IRC12:03
*** sapcc-bot1 has quit IRC12:03
*** sapcc-bot has joined #openstack-manila12:04
*** tpatzig_ has joined #openstack-manila12:04
*** carthaca_ has joined #openstack-manila12:04
*** seife_ has joined #openstack-manila12:04
*** databus23_ has joined #openstack-manila12:04
tbarrontommylikehu: pong12:04
*** seife_ has quit IRC12:06
*** tpatzig_ has quit IRC12:06
*** databus23_ has quit IRC12:06
*** tommylikehu_ has quit IRC12:06
*** ociuhandu has joined #openstack-manila12:13
arnewiebalckI’ve put together a brief spec to allow for quotas per share type: https://review.openstack.org/#/c/447021/. Please have a look and let me know if you have comments or questions.12:14
*** tommylikehu_ has joined #openstack-manila12:15
*** tommylikehu_ has quit IRC12:16
*** tommylikehu_ has joined #openstack-manila12:17
tommylikehu_Thanks tbarron, I think you are one of the people who interested in the user message which I had talked with jprovazn a few hours ago12:18
openstackgerritArne Wiebalck proposed openstack/manila-specs master: Add spec for per share type quotas  https://review.openstack.org/44702112:18
* tbarron reads backlog12:19
openstackgerritArne Wiebalck proposed openstack/manila-specs master: Add spec for per share type quotas  https://review.openstack.org/44702112:19
tbarrontommylikehu_: jprovazn good conversation.  Tommy: looks like you are going to take your ideas on that front to the cinder meetingi today?12:20
tbarrontommylikehu_: meeting12:20
tbarrontommylikehu_: it does seem to me that sharing as much of this infra between cinder and manila is a good idea.12:20
tommylikehu_yes12:21
tommylikehu_sheer is not working in cinder any more12:21
tbarrontommylikehu_: jprovazn I think for cinder people are working on exposing these messages in horizon, so we can see how that goes and follow in manila (if it goes well)12:21
tommylikehu_tbarron: hope it could goes well12:22
tbarrontommylikehu_: speaking of cinder/manila overlaps, note arnewiebalck that he mentioned a moment ago.12:22
tbarrontommylikehu_: cinder already has per-volume-type quotas12:22
tbarronarnewiebalck: tommylikehu_ this area is complicated a little bit by:12:23
tbarron1) cinder has hierarchical quotas, manila doesn't12:23
tbarron2) manila has per-user quotas, cinder doesn't12:23
tbarronarnewiebalck: does CERN use or care about per-user quotas or hierarchical quotas?12:25
tbarronvs regular old tenant/project quotas?12:26
arnewiebalcktbarron: yes, we’re pushing for hierarchical quotas12:26
smcginnisPer-user is just plain insanity.12:26
tbarronarnewiebalck: oh yeah, I see tim bell as a reviewer on the keystone limits spec12:26
tommylikehu_arnewiebalck: you come from cern?12:26
tbarronsmcginnis: don't disagree, but manila and neutron have it12:27
arnewiebalcktommylikehu_: yes12:27
smcginnistbarron: Manila and neutron are just plain insanity.12:27
smcginnis:)12:27
tbarronsmcginnis: i've had to learn a lot more neutron working on manila rather than cinder, so you be the judge :)12:27
tommylikehu_smcginnis: it's nice you are here, tbarron and I just talked about the user message stuff12:28
arnewiebalcktbarron: the lack of hierarchical quotas makes it more complicated as you cannot easily use Cinder’s implementation as a guideline or because you you need hierarchical quotas?12:28
tommylikehu_smcginnis: maybe you could take a look the commit message of this patch: https://review.openstack.org/#/c/448384/12:28
tommylikehu_and leave your concerns there :)12:29
smcginnistommylikehu_: Sure, I'll try to get to that.12:29
tbarronarnewiebalck: I think it's actually just a minor wrinkle for porting per-volume-quotas over as per-share-quotas, not a big deal.12:29
arnewiebalcktbarron: I had a look at the Cinder patch referenced in the spec12:31
arnewiebalcktabrron: didn’t look too complicated to me, but I wouldn’t know if it relies on things that do not exist in Manila12:31
tbarronarnewiebalck: I think it would be pretty straightforward but it has been six months or so since I actually looked at it in detail.12:32
tbarronarnewiebalck: tommylikehu_ also relevant for anything we do new with quotas: tommylikehu_: arnewiebalck also relevant is https://review.openstack.org/44081512:33
tbarronkeystone would hold limit information, projects would still actually enforce quotas12:33
tbarronarnewiebalck: tommylikehu_ This is more for when/if we implement hierarchical quotas.12:34
*** markstur has joined #openstack-manila12:34
tbarronarnewiebalck: tommylikehu_ I don't think it would impact per-share-quotas in any big way, mostly just something to be aware of and to make sure we don't do something that makes using this keystone-based-limits stuff harder.12:35
tbarronsmcginnis: i'd expect that if 440815 moves ahead (and it seems to have support) cinder hierarchical quotas would eventually12:36
tbarronget re-worked to use it, and neutron and manila could drop per-user quotas in favor of a common12:36
tbarronapproach to hierarchical quotas.12:36
smcginnistbarron: I hope so.12:36
smcginnistbarron: I would much rather have a common solution than each of us with our own quirks.12:37
tbarronsmcginnis: +112:37
tommylikehu_tbarron, does the quota works well without keystone if we finally have this feature?12:37
*** cknight has joined #openstack-manila12:38
tbarrontommylikehu_: that's a great question since as we have discussed running cinder/manila as SDS w/o the rest of OpenStack is a goal too12:38
tbarrontommylikehu_: in your free time :) maybe check that review and make sure that concern is dealt with/addressed.12:39
*** markstur has quit IRC12:39
tommylikehu_tbarron: thanks for your suggestion. I will do as soon as I take up my English dictionary12:40
tbarrontommylikehu_: I think you could write that dictionary :)12:41
*** catintheroof has joined #openstack-manila12:42
*** xyang1 has joined #openstack-manila12:43
arnewiebalcktbarron: FWIW, I just checked the status of hierarchical quotas in nova: it has a spec for which information is gathered; seems to be somewhat away still.12:46
*** markstur has joined #openstack-manila12:59
*** dsariel_ has quit IRC13:01
*** markstur has quit IRC13:03
tbarronarnewiebalck: yep, cinder is the leader on that front, but for manila I think it makes sense to also track nova and the keystone-limit thing rather than just porting over what cinder has.13:05
tbarronarnewiebalck: that means hierarchical quotas will be a bit further out ...13:05
arnewiebalcktbarron: absolutely13:05
tbarronarnewiebalck: but the per-share-type stuff can happen sooner if we can get someone to do the coding.13:05
tbarronand testing13:05
arnewiebalcktbarron: tesing should be covered ;)13:06
arnewiebalcks/tesing/testing/13:06
tbarronarnewiebalck: yeah, I guess deploying at CERN would give us a bit of coverage13:06
tbarronarnewiebalck: need tempest tests written/adjusted too though13:06
tbarrons/bit of coverage/bit of coverage :)/13:07
arnewiebalcktbarron :)13:08
*** eharney has joined #openstack-manila13:13
*** dsariel_ has joined #openstack-manila13:14
*** carthaca_ has quit IRC13:19
*** sapcc-bot has quit IRC13:19
*** carthaca_ has joined #openstack-manila13:19
*** sapcc-bot has joined #openstack-manila13:19
*** gouthamr has joined #openstack-manila13:21
*** dustins has joined #openstack-manila13:41
*** crushil has joined #openstack-manila13:48
*** mtanino has joined #openstack-manila13:50
*** tongli has joined #openstack-manila13:59
*** chlong has joined #openstack-manila14:04
*** chlong has quit IRC14:12
*** chlong has joined #openstack-manila14:13
*** eharney has quit IRC14:15
*** Yogi1 has joined #openstack-manila14:16
openstackgerritArne Wiebalck proposed openstack/manila-specs master: Add spec for per share type quotas  https://review.openstack.org/44702114:32
*** markstur has joined #openstack-manila14:37
*** crushil has quit IRC14:38
*** markstur has quit IRC14:41
*** eharney has joined #openstack-manila14:51
*** tongli has quit IRC15:01
*** xyang_ has joined #openstack-manila15:05
*** tommylikehu_ has quit IRC15:07
*** tommylikehu_ has joined #openstack-manila15:09
*** mtanino has quit IRC15:10
*** markstur has joined #openstack-manila15:11
*** mtanino has joined #openstack-manila15:13
*** mkoderer has joined #openstack-manila15:27
*** chlong has quit IRC15:30
*** chlong has joined #openstack-manila15:33
*** akapil has joined #openstack-manila15:37
*** akapil has quit IRC15:40
*** xyang_ has quit IRC15:54
*** xyang_ has joined #openstack-manila15:55
*** xyang_ has quit IRC15:59
*** xyang_ has joined #openstack-manila16:00
*** nkrinner is now known as nkrinner_afk16:05
*** harlowja has joined #openstack-manila16:06
*** kaisers_ has joined #openstack-manila16:14
*** ganso has joined #openstack-manila16:15
*** gouthamr-afk has joined #openstack-manila16:20
*** gouthamr_afk has joined #openstack-manila16:26
*** gouthamr-afk has quit IRC16:26
*** gouthamr-afk has joined #openstack-manila16:28
*** rraja has quit IRC16:29
*** gouthamr_afk has quit IRC16:31
*** chlong has quit IRC16:45
*** chlong has joined #openstack-manila16:52
*** xyang_ has quit IRC17:16
*** xyang_ has joined #openstack-manila17:16
*** tommylikehu_ has quit IRC17:20
*** xyang_ has quit IRC17:23
*** kaisers__ has joined #openstack-manila17:24
*** kaisers_ has quit IRC17:27
*** gouthamr-afk has quit IRC17:28
*** gouthamr_afk has joined #openstack-manila17:28
*** xyang_ has joined #openstack-manila17:29
*** xyang_ has quit IRC17:30
*** swapnil-nilangek has joined #openstack-manila17:32
*** mtanino has quit IRC17:33
*** ociuhandu has quit IRC17:34
*** kaisers has quit IRC17:41
*** ociuhandu has joined #openstack-manila17:52
*** kaisers has joined #openstack-manila17:56
*** ociuhandu has quit IRC17:58
*** ociuhandu has joined #openstack-manila18:02
*** ociuhandu has quit IRC18:07
*** swapnil-nilangek has quit IRC18:12
*** mtanino has joined #openstack-manila18:12
*** gouthamr_afk has quit IRC18:14
openstackgerritValeriy Ponomaryov proposed openstack/manila master: Add possibility to run 'manila-api' with wsgi web servers  https://review.openstack.org/44819018:30
vponomaryovbswartz: ^18:30
openstackgerritValeriy Ponomaryov proposed openstack/manila master: Add possibility to run 'manila-api' with wsgi web servers  https://review.openstack.org/44819018:35
*** kaisers has quit IRC18:40
*** xyang_ has joined #openstack-manila18:45
*** xyang_ has quit IRC18:46
*** kaisers has joined #openstack-manila18:56
*** harlowja has quit IRC18:59
*** harlowja has joined #openstack-manila19:00
*** harlowja has quit IRC19:01
*** harlowja has joined #openstack-manila19:45
*** harlowja has quit IRC20:18
*** kaisers has quit IRC20:20
*** Yogi1 has quit IRC20:32
*** kaisers has joined #openstack-manila20:35
*** crushil has joined #openstack-manila20:40
*** crushil has left #openstack-manila20:43
*** eharney has quit IRC20:46
*** mkoderer has quit IRC20:47
*** cknight has quit IRC20:52
*** kaisers has quit IRC21:06
*** chlong has quit IRC21:08
*** dustins has quit IRC21:21
*** kaisers has joined #openstack-manila21:21
*** jprovazn has quit IRC21:23
*** gouthamr has quit IRC21:26
*** kaisers__ has quit IRC21:30
*** harlowja has joined #openstack-manila21:33
*** pcaruana has quit IRC21:47
*** kaisers has quit IRC21:59
*** kaisers has joined #openstack-manila22:13
*** kaisers_ has joined #openstack-manila22:31
*** kaisers has quit IRC22:33
*** kaisers_ has quit IRC22:36
*** xyang1 has quit IRC22:38
*** kaisers has joined #openstack-manila22:48
*** sticker has quit IRC22:55
*** sticker has joined #openstack-manila22:55
*** catintheroof has quit IRC23:01
*** catintheroof has joined #openstack-manila23:02
*** catintheroof has quit IRC23:06
*** tommylikehu_ has joined #openstack-manila23:12
*** tommylikehu_ has quit IRC23:13
*** ganso has quit IRC23:25
*** kaisers_ has joined #openstack-manila23:32
*** kaisers_ has quit IRC23:37
*** mtanino has quit IRC23:37

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!